PREDICTION OF THE DIGESTIBLE ENERGY AND DIGESTIBILITY OF GROSS ENERGY OF FEEDS FOR RABBITS .2. MIXED DIETS

Mixed diets, for which information relating to various chemical measurements and rates of inclusion of certain raw materials (independent variables) and their nutritive value for rabbits (dependent variables) as assessed in terms of digestible energy (DE) and the coefficient of digestibility of gross energy (GE(D)) was available, were combined in a series of correlation matrices and step-wise linear regression analyses. GE(D) removes the possible effect of variability in gross energy value on DE values and is reported. In general, GE(D) was better correlated to acid detergent fibre (ADF) than crude fibre (CF). A general equation was derived: GE(D) = 0.867 - 0.0012 ADF (g kg-1 DM) R2 = 0.888. Closer inspection of individual points revealed that diets based on added fat, beet pulp, citrus pulp and straw at levels higher than 200 g kg-1 behaved atypically. Separate prediction equations were derived for each of these groups of diets. Multiple regression analyses considered more than one independent variable, but gave only marginal improvements in terms of the accuracy of prediction of GE(D). Addition of rates of inclusion of feedstuffs as independent variables revealed that the chemical analyses employed were insensitive to changes in nutritive value arising from additions of fat, pulps and straw. These feedstuffs need to be characterised further and general prediction equations should not be applied to diets in which they are contained.